Research On The Acquisition Of Interdisciplinary Data Collection In Humanities And Social Sciences Based On The Fusion Of Domain Knowledge Graph

With the deepening of knowledge graph research and application,knowledge sharing and fusion of multi-source heterogeneous interdisciplinary knowledge graphs become a new demand.The interdisciplinarity of disciplines is increasing year by year,and the requirements of data sets for each field are more in need of graphs with professional attributes,i.e.,Domain Knowledge Graph（DKG）.Through knowledge extraction of the names of the National Social Science Foundation（NSSF）projects from 2016-2022,23 hot disciplines of social science in the last 7 years are derived,and then through statistical literature analysis of the NSSF papers,it is also found that education has a high degree of intersection with multiple disciplines,so this thesis also includes education into the study.The coupling strength coefficient of disciplinary intersection was calculated by extracting the high-frequency keywords from the list of projects of the National Social Science Foundation of China in the past 7 years,forming a network density graph of disciplinary intersection,using the top 5%of keywords among 24 disciplines to determine the disciplinary intersection candidate set to generate the discipline-keyword DKG,and deriving the annual characteristics and 7-year overall regularity of disciplinary intersection in social sciences.Deep mining the binary disciplinary intersection（education-management）graph,improving the small-sample graph fusion algorithm-VCU system（2+i,layer i is the acquisition graph layer）,improved Wup K_i,Key Match,F₁formulas,graphping screening and fusion of binary disciplinary intersection domain graphping with two candidate sets;then quality check of binary disciplinary intersection domain knowledge graphping using new evaluation metrics;also the matching degree of two candidate sets can be evaluated;then the candidate sets are preferentially selected for graphping fusion.The preliminary One-Dimensional Collect Knowledge Graph of Interdisciplinary Areas（ODCKGIA）is generated.Initially generated ODCKGIA was subjected to ontology summary clustering to generate candidate major and minor subsets,and merged into the 1st ontology clustering set of ODCKGIA,filtered using improved semantic filtering formulas（improved local Density,global Density,local Key,global Density,NCvalue）The Public Topics（Pub T）and Potential Topics（Pot T）were generated,and then the Semantic Text Filtering Importance Measurement Model（STFIMM）was constructed to empirically analyze the construction of STFIMM with"Teacher+Teaching+Innovation"as the main predicate.The core triad of"teacher+teaching+innovation"as the main predicate was searched by keywords in Questionnaire Star,and 13 collected semantic texts were searched to match the semantic filtering of"teacher+teaching+innovation"+ODCKGIA,and new Pub T’and Pot T’were generated,construct the Collect Private Knowledge Graph of Interdisciplinary Areas（CPKGIA）,and perform cluster analysis on the Collect Private Graph to generate the semantic texts belonging to the collection.The generated collection semantic texts are verified to be effective using reliability and SEM path evolution analysis.In summary,this thesis improves the small-scale domain graphping fusion algorithm to construct ODCKGIA and CPKGIA,which provides a reference for the subsequent interdisciplinary research of humanities and social science disciplines and collection data.